Proxmox 6.2 / Corosync 3 - rare and spontaneous disruptive UDP:5405-storm/flood

[Aminuxer]

Nowtime Proxmox updated to 6.4
But our two smaller clusters working separated - we avoid build one big cluster again.

Does exists any news/comments about this bug ? Does it really fixed in 6.3 / 6.4 version ?

 

[spirit]

Im' not sure it's fixed in 6.4.

"maybe" in last 7.X corosync package.

with this fix:

https://github.com/corosync/corosync/commit/cdf72925db5a81e546ca8e8d7d8291ee1fc77be4

 

[IrisDande]

@spirit: the corosync package in your link above is released at corosync version 3.1.6 but on pve host still is 3.1.5-pve2.
May I know when the packages will be released for 3.1.6-pve2?

root@pve1:~# dpkg -l | grep corosync
ii  corosync                             3.1.5-pve2                     amd64        cluster engine daemon and utilities
ii  libcorosync-common4:amd64            3.1.5-pve2                     amd64        cluster engine common library

 

[fabian]

3.1.5-pve2 contains the relevant fixes (we work closely with the corosync upstream, and pulled them in before they cut their 3.1.6 release )

https://git.proxmox.com/?p=corosync-pve.git;a=commit;h=ee66dad2ce924fe6a31496c043ddeeedfa098b98

 

[IrisDande]

Weird, I have 15 node with split brain now, Can I continues on this thread or open a new one

 

[fabian]

please take a look at the logs (corosync and pve-cluster mainly) - but yeah, opening a new thread is probably a good idea.